1. Field of the Invention
This invention relates to a code conversion apparatus producing run-length limited codes suitable for recording digital signals.
2. Description of the Prior Art
In case where digital data are recorded in magnetic tapes or discs at high densities, run-length limited codes (referred to as RLL codes hereinafter) are generally used. In an RLL code system, m-bit data words are converted into respective n-bit code words, which are concatenated to form a bit sequence where the number of successive bits having the same binary value is limited to a range of a smaller value d to a larger value k.
RLL codes suitable for high density recording are generally required to satisfy the following three known conditions.
[1] The detection window width Tw (=m/n.multidot.T) is wide, where the letter T denotes the one-bit length of a data word. PA0 [2] The minimal inversion interval Tmin (=d.multidot.Tw) is large. PA0 [3] The maximal successive bit number k is small.
As the detection window width Tw increases, variations in the time axis, such as jitters or peak shifts, less affect the rate of errors in coding during reproduction or playback process. As the minimal inversion interval Tmin increases, high-frequency component cut-off characteristics of recording and reproducing systems less affect qualities of reproduced digital signals. As the maximal successive bit number k decreases, it is easier to obtain self clock function deriving clock information from a reproduced or playback signal. Since the detection window width and the minimal inversion interval greatly affect recording and reproducing characteristics, they are important.
Heretofore, various RLL codes have been developed. U.S. Pat. No. 3,689,899 discloses 2/3 codes which are one type of RLL codes. In these 2/3 codes, the values d, k, and Tw are equal to 2, 9, and 0.667 T respectively. The 2/3 codes constitute variable length RLL codes which change between a first state where 2-bit data words are converted into 3-bit code words and a second state where 4-bit data words are converted into 6-bit code words.
Although the 2/3 codes satisfy the previously-mentioned desired conditions [1]-[3], the 2/3 codes have problems as follows. The 2/3 codes use both the 2/3 conversion and the 4/6 conversion. Therefore, during the code conversion of 8-bit digital video data, a boundary or pause between data words sometimes extends over two words of video data. In this case, if an error occurs in a code word, this error spreads or expands over the two words of the video data during decoding. It should be noted that one word of video data has 8 bits.
Although the 2/3 RLL codes are used to prevent an increase in the rate of decoding errors in high density recording, the use of the 2/3 codes cause expansions of decoding errors in some cases since the 2/3 codes are of the variable length. Especially, in the case of video tape recorders for personal or family use, the density of recording is generally set high to allow a long recording. Accordingly, decoding errors are liable to occur during play back process, and expansions of errors frequently occur when the 2/3 RLL codes are used.
U.S. Pat. No. 3,852,687 discloses 8/13 codes which have the largest detection window width among conventional RLL codes allowing direct code conversion of 8-bit data words and having the value d equal to 2. The detection window width of the 8/13 codes equals 0.615 T and is smaller than the detection window width of the 2/3 codes. In general, the smaller detection window width deteriorates the recording and reproducing characteristics and also increases the error rate.